Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Ontological engineering for rhetorical figures and
beyond
Cliff O’Reilly
Computational Rhetoric Workshop
Waterloo, 13th August 2016
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Outline
1 Ontological Engineering
2 Linking Ontologies
3 Lassoing Rhetoric
4 Current work
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Lassoing Rhetoric
Current work
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Ontological Engineering
Expressivity
There are infinite ways to model the world
There are restrictions for a given framework, e.g. types and
properties
It’s a creative exercise
There are no wrong answers (until the model breaks)
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Discussion
Ontological Engineering
Architecture
OWL/RDFS due to its interoperability and expressivity
Key distinction between the ontology (e.g. owl file) and the
knowledge base (e.g. RDF triples)
Deliver the knowledge base and ontology together
Use an RDF Store to separate ontology from Knowledge Base
Thinking about usage is important
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Ontological Engineering
Architecture - example
Anaphora is a Rhetorical Device
Expressed as an RDF triple in XML, e.g.
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Current work
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Linking Ontologies
Why?
Easy to do
Functional separation
Powerful, e.g. increased scope for inference
All elements can be accessed from the linked ontology
Semantic Web principles - opportunity for targeted re-use
Can become unwieldy
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Discussion
Linking Ontologies
How?
Linking ontologies
W3C standard gives the owl:imports element which is added to the
header in the containing OWL ontology and pointed to the
resource to import, e.g. <owl:imports rdf:resource=”http:
//repositori.com/sw/onto/RhetoricalDevices.owl”/>
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Lassoing Rhetoric
Background
2010 MSc dissertation project
Automatic rhetorical figure detection using Semantic Web
technology
OWL ontologies and rules in SWRL (Semantic Web Rule
Language)
A personal interest in rhetorical patterns and history
Unpublished paper: ”Lassoing Rhetoric with OWL and
SWRL”
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Discussion
Lassoing Rhetoric
The model
Java program, calling on GATE1 , Protege2 and JESS3
reasoner APIs
Suite of OWL ontologies
Extension into Semantic Web - Linked Data
Designed for inference - discovering new knowledge via logic
rules
Ontologies designed with the goal in mind, i.e. reverse
engineered from logical rules
1
https://gate.ac.uk/
http://protegewiki.stanford.edu/wiki/ProtegeOWL API Programmers Guide
3
http://www.jessrules.com/jess/index.shtml
2
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Lassoing Rhetoric
In detail - ontologies
Controller ontology
Lassoing Rhetoric (http:
//repositori.com/sw/onto/LassoingRhetoric.owl)
Imported ontologies
Document Structure
(http://repositori.com/sw/onto/DocStruct.owl)
Rhetorical Devices (http:
//repositori.com/sw/onto/RhetoricalDevices.owl)
Rhetorical Structures (http:
//repositori.com/sw/onto/RhetoricalStruct.owl)
Language Tags
(http://repositori.com/sw/onto/langtag.owl)
GATE (http://repositori.com/sw/onto/gate.owl)
Verb Noun Combinations
(http://repositori.com/sw/onto/VerbNounCombo.owl)
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Lassoing Rhetoric
In detail - example 1
Anaphora (Epanaphora)
A figure characterised by a repeating word at the beginning of
successive clauses, e.g. Churchill’s speech to Parliament in 1940:
”We shall go on to the end. We shall fight in France. We shall
fight on the seas and oceans..”.
The pattern chosen to be modelled in this project is a two word
phrase repeating in successive sentence beginnings.
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Lassoing Rhetoric
In detail - example 1
Anaphora (Epanaphora) - example 1 - simplified syllogism
IF Document HAS Paragraph
AND IF Paragraph HAS Sentence
AND IF Sentence HAS Clause A
AND IF Sentence HAS Clause B
AND IF Clause A HAS Word X
AND IF Clause A HAS Word Y
AND IF Clause B HAS Word F
AND IF Clause B HAS Word G
AND IF Word X IS THE SAME AS Word F
AND IF Word Y IS THE SAME AS Word G
THEN Document HAS Anaphora
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Discussion
Lassoing Rhetoric
In detail - example 1
Anaphora (Epanaphora) - Horn Clause
DocStruct:Doc(?h) ∧ DocStruct:hasParagraph(?h,?i)
∧ gate:Sentence(?z) ∧ gate:Sentence(?c)
∧ gate:word(?a) ∧ gate:word(?x)
∧ DocStruct:hasSentence(?i,?z) ∧ DocStruct:hasNextSentence(?z,?c)
∧ DocStruct:hasFirstWord(?z,?x) ∧ DocStruct:hasFirstWord(?c,?a)
∧ DocStruct:hasNextWord(?x,?y) ∧ DocStruct:hasNextWord(?a,?b)
∧ gate:hasString(?x,?d) ∧ gate:hasString(?y,?e) ∧ gate:hasString(?a,?f) ∧
gate:hasString(?b,?g)
∧ swrlb:equal(?d,?f) ∧ swrlb:equal(?e,?g)
∧ gate:hasStartNode(?x,?j) ∧ gate:hasEndNode(?b,?k)
=⇒ RhetDev : hasRhetoricalDevice(?h, RhetDev : Anaphora)∧
gate:hasStartNode(RhetDev:Anaphora,?j) ∧ gate:hasEndNode(RhetDev:Anaphora,?k)
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Discussion
Lassoing Rhetoric
In detail - example 2
Epizeuxis
A figure characterised by an emphatic repetition of a word with no
other words in between, e.g. Poe’s poem The Bells:
”How they tinkle, tinkle, tinkle ... From the bells, bells, bells, bells
...”.
The pattern chosen to be modelled in this project is the same word
repeating three times in succession.
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Discussion
Lassoing Rhetoric
In detail - example 2
Epizeuxis - Horn Clause
DocStruct:Doc(?d) ∧ DocStruct:hasNextWord(?x, ?y) ∧ DocStruct:hasNextWord(?y,
?z)
∧ gate:hasString(?x, ?a) ∧ gate:hasString(?y, ?b) ∧ gate:hasString(?z, ?c)
∧ swrlb:equal(?a, ?b) ∧ swrlb:equal(?a, ?c)
∧ gate:hasStartNode(?x, ?e) ∧ gate:hasEndNode(?z, ?f)
=⇒ RhetDev : hasRhetoricalDevice(?d, RhetDev : Epizeuxis)∧
gate:hasStartNode(RhetDev:Epizeuxis, ?e) ∧ gate:hasEndNode(RhetDev:Epizeuxis, ?f)
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Discussion
Lassoing Rhetoric
In detail - extension
A key idea from the project was to take advantage of Linked
Data, e.g.
DBPedia -utilising data from the web to inform the
ontology/knowledge base, such as world knowledge
Wordnet - using data to extend the knowledge base, e.g. by
including synonyms and resultant inferences
Facilitated via the Java program, but dynamic and
parameterised
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Discussion
Lassoing Rhetoric
In detail - example 3
Historic Present
A figure characterised by a reference to a thing in the past as if it
were still referable in the present, such as a person no longer alive
having thoughts in the present day. For example,
”What Socrates is trying to do here is...”.
The pattern chosen to be modelled in this project is based on the
instantiation of a class called DeadPerson and this being
referenced by a verb in the present tense.
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Lassoing Rhetoric
In detail - example 3
Historic Present - Horn Clause
DocStruct:Doc(?x) ∧ gate:paragraph(?y) ∧ DocStruct:hasParagraph(?x, ?y)
∧ gate:Sentence(?z) ∧ DocStruct:hasSentence(?y, ?z)
∧ gate:word(?a) ∧ DocStruct:hasWord(?z, ?a)
∧ gate:word(?b) ∧ DocStruct:hasWord(?z, ?b)
∧ DeadPerson(?a) ∧ langtag:ThirdPersonSingularPresentVerb(?b) ∧
gate:hasDependency(?b, ?a)
∧ gate:hasStartNode(?a, ?c) ∧ gate:hasEndNode(?a, ?d)
=⇒ RhetDev : hasRhetoricalDevice(?x, RhetDev : Historic Present)∧
gate:hasStartNode(RhetDev:Historic Present, ?c) ∧
gate:hasEndNode(RhetDev:Historic Present, ?d)
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Current work
Ontology of cognition and intelligence
A new ontology to capture primitive cognitive patterns
Early stages of development
PhD proposal
Domain: Intelligent agents in 3D gaming
... however, it could apply to any application especially
analysis of Rhetoric
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Current work
Ontology of cognition and intelligence
Influence from Cognitive Science, Cognitive Linguistics and
Embodiment theory
Currently the most primitive aspects of cognition are under
development, e.g.
Identity / Categorisation
Perception of primitive shapes, e.g. lines and boundaries etc
Material behaviour
Quantity / number
Persistence / temporality
Size, Scale etc
Proximity
Difference / similarity
The scope includes developing more complex functions
involving intentionality, goals, decisions, beliefs, frames etc
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Current work
Cognitive Linguistics and Embodiment
Image Schema; Conceptual Metaphor - Lakoff and Johnson
Containment, location, motion, lines, adjacency, contact,
orientation etc
Metaphors - More-Is-Up, Less-Is-Down, Love-As-Journey etc
Embodiment; Situated/Embodied Cognition
Cognition is grounded in physical aspects and implications of
being in a body in the world
Natural language, and therefore Rhetorical forms, is strongly
influenced by this underlying cognitive framework, e.g. via
metaphors and words such as Up/Down etc
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Discussion
Ontological Engineering
Linking Ontologies
Lassoing Rhetoric
Current work
Discussion
Current work
How does this relate to rhetoric?
Descriptive, e.g. adding more knowledge to our models
New ways of describing patterns
Discovering new knowledge via inference
”Ask not what your country can do for you; ask what you can
do for your country” (Antimetable) could be described in
terms of:
Relative proximity (physical and/or other) of like-terms
”Actions are self-propelled movements” - Event Structure
conceptual metaphor
Size comparison of what a country can do versus what you can
do
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Ontological Engineering
Linking Ontologies
Discussion
Questions?
Suggestions?
Comments?
Cliff O’Reilly
Ontological engineering for rhetorical figures and beyond
Lassoing Rhetoric
Current work
Discussion